Hebrew - Wikilangs Models
Comprehensive Research Report & Full Ablation Study
This repository contains NLP models trained and evaluated by Wikilangs, specifically on Hebrew Wikipedia data. We analyze tokenizers, n-gram models, Markov chains, vocabulary statistics, and word embeddings.
π Repository Contents
Models & Assets
- Tokenizers (8k, 16k, 32k, 64k)
- N-gram models (2, 3, 4, 5-gram)
- Markov chains (context of 1, 2, 3, 4 and 5)
- Subword N-gram and Markov chains
- Embeddings in various sizes and dimensions (aligned and unaligned)
- Language Vocabulary
- Language Statistics
Analysis and Evaluation
- 1. Tokenizer Evaluation
- 2. N-gram Model Evaluation
- 3. Markov Chain Evaluation
- 4. Vocabulary Analysis
- 5. Word Embeddings Evaluation
- 6. Morphological Analysis (Experimental)
- 7. Summary & Recommendations
- Metrics Glossary
- Visualizations Index
1. Tokenizer Evaluation
Results
| Vocab Size | Compression | Avg Token Len | UNK Rate | Total Tokens |
|---|---|---|---|---|
| 8k | 3.129x | 3.13 | 0.0482% | 4,188,199 |
| 16k | 3.502x | 3.50 | 0.0540% | 3,742,094 |
| 32k | 3.872x | 3.87 | 0.0597% | 3,384,734 |
| 64k | 4.191x π | 4.19 | 0.0646% | 3,127,199 |
Tokenization Examples
Below are sample sentences tokenized with each vocabulary size:
Sample 1: ΧΧΧΧΧ Χ©ΧΧΧΧ ΧΧ ΧΧΧΧ Χ©ΧΧΧ (Eisenstein), Χ©Χ ΧΧ©Χ€ΧΧ ΧΧ¨ΧΧ Χ ΧΧ©Χ ΧΧΧΧΧ ΧΧ©ΧΧ ΧΧ Χ Χ€ΧΧ₯. Χ€ΧΧ¨ΧΧ©...
| Vocab | Tokens | Count |
|---|---|---|
| 8k | βΧΧΧΧ Χ Χ©ΧΧΧΧ βΧΧ βΧΧΧ Χ Χ©Χ ΧΧ β( e is ... (+26 more) |
36 |
| 16k | βΧΧΧΧ Χ Χ©ΧΧΧΧ βΧΧ βΧΧΧ Χ Χ©Χ ΧΧ β( e is en ... (+20 more) |
30 |
| 32k | βΧΧΧΧ Χ Χ©ΧΧΧΧ βΧΧ βΧΧΧ Χ Χ©Χ ΧΧ β( e is en ... (+19 more) |
29 |
| 64k | βΧΧΧΧΧ Χ©ΧΧΧΧ βΧΧ βΧΧΧ Χ Χ©Χ ΧΧ β( e is enstein ), ... (+17 more) |
27 |
Sample 2: Χ©ΧΧΧΧ ΧΧΧ Χ¦ΧΧ¨Χͺ ΧΧ§ΧΧ Χ Χ©Χ ΧΧΧΧΧ ΧΧΧΧΧΧͺ Χ©ΧΧΧ ("ΧΧΧͺ" ΧΧ "ΧΧΧ¨"). ΧΧ©Χ€ΧΧ ΧΧ©Χ€ΧΧ ΧΧ©ΧΧ ΧΧΧΧ
| Vocab | Tokens | Count |
|---|---|---|
| 8k | βΧ©Χ ΧΧΧ βΧΧΧ βΧ¦ΧΧ¨Χͺ βΧΧ§ΧΧ Χ βΧ©Χ βΧΧΧΧΧ βΧΧ ΧΧΧΧͺ βΧ©Χ ... (+13 more) |
23 |
| 16k | βΧ©Χ ΧΧΧ βΧΧΧ βΧ¦ΧΧ¨Χͺ βΧΧ§ΧΧ Χ βΧ©Χ βΧΧΧΧΧ βΧΧ ΧΧΧΧͺ βΧ©Χ ... (+12 more) |
22 |
| 32k | βΧ©Χ ΧΧΧ βΧΧΧ βΧ¦ΧΧ¨Χͺ βΧΧ§ΧΧ Χ βΧ©Χ βΧΧΧΧΧ βΧΧ ΧΧΧΧͺ βΧ©Χ ... (+11 more) |
21 |
| 64k | βΧ©ΧΧΧΧ βΧΧΧ βΧ¦ΧΧ¨Χͺ βΧΧ§ΧΧ Χ βΧ©Χ βΧΧΧΧΧ βΧΧΧΧΧΧͺ βΧ©Χ ΧΧ β(" ... (+9 more) |
19 |
Sample 3: ΧΧΧΧ€Χ¨Χ ΧΧΧ ΧΧͺΧ’ΧͺΧΧ§ ΧΧ’ΧΧ¨Χ ΧΧΧΧΧ Leopard, ΧΧ§ΧΧΧΧͺ ΧΧΧ‘Χ€Χ¨ Χ©Χ€ΧΧͺ ΧΧΧ©ΧΧ’ΧΧͺΧ ΧΧΧ Χ ΧΧ¨ (ΧΧ’Χ Χ...
| Vocab | Tokens | Count |
|---|---|---|
| 8k | βΧΧ ΧΧ€Χ¨ Χ βΧΧΧ βΧΧͺ Χ’ΧͺΧΧ§ βΧΧ’ΧΧ¨Χ βΧΧ ΧΧΧ βle ... (+21 more) |
31 |
| 16k | βΧΧ ΧΧ€Χ¨ Χ βΧΧΧ βΧΧͺ Χ’ΧͺΧΧ§ βΧΧ’ΧΧ¨Χ βΧΧΧΧΧ βle op ... (+17 more) |
27 |
| 32k | βΧΧΧΧ€Χ¨ Χ βΧΧΧ βΧΧͺ Χ’ΧͺΧΧ§ βΧΧ’ΧΧ¨Χ βΧΧΧΧΧ βle op ard ... (+15 more) |
25 |
| 64k | βΧΧΧΧ€Χ¨ Χ βΧΧΧ βΧΧͺΧ’ΧͺΧΧ§ βΧΧ’ΧΧ¨Χ βΧΧΧΧΧ βle opard , βΧΧ§ΧΧΧΧͺ ... (+12 more) |
22 |
Key Findings
- Best Compression: 64k achieves 4.191x compression
- Lowest UNK Rate: 8k with 0.0482% unknown tokens
- Trade-off: Larger vocabularies improve compression but increase model size
- Recommendation: 32k vocabulary provides optimal balance for production use
2. N-gram Model Evaluation
Results
| N-gram | Variant | Perplexity | Entropy | Unique N-grams | Top-100 Coverage | Top-1000 Coverage |
|---|---|---|---|---|---|---|
| 2-gram | Word | 839,907 | 19.68 | 4,883,996 | 3.8% | 9.8% |
| 2-gram | Subword | 388 π | 8.60 | 45,811 | 57.3% | 98.0% |
| 3-gram | Word | 2,460,970 | 21.23 | 7,456,944 | 1.9% | 5.1% |
| 3-gram | Subword | 4,159 | 12.02 | 320,573 | 19.8% | 57.8% |
| 4-gram | Word | 6,086,424 | 22.54 | 12,242,689 | 1.3% | 3.3% |
| 4-gram | Subword | 31,153 | 14.93 | 1,768,539 | 7.8% | 25.6% |
| 5-gram | Word | 5,115,710 | 22.29 | 8,563,842 | 1.1% | 3.0% |
| 5-gram | Subword | 174,825 | 17.42 | 6,204,970 | 3.7% | 13.2% |
Top 5 N-grams by Size
2-grams (Word):
| Rank | N-gram | Count |
|---|---|---|
| 1 | Χ’Χ ΧΧΧ |
619,385 |
| 2 | Χ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ |
326,599 |
| 3 | ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ |
252,301 |
| 4 | ΧΧ¨Χ¦ΧΧͺ ΧΧΧ¨ΧΧͺ |
176,732 |
| 5 | Χ’Χ Χ€Χ |
148,464 |
3-grams (Word):
| Rank | N-gram | Count |
|---|---|---|
| 1 | Χ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ |
115,186 |
| 2 | ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ |
115,178 |
| 3 | Χ©Χ ΧΧ¨Χ¦ΧΧͺ ΧΧΧ¨ΧΧͺ |
67,555 |
| 4 | Χ©Χ ΧΧΧΧ Χ |
45,554 |
| 5 | ΧΧΧΧ Χ 20 |
39,531 |
4-grams (Word):
| Rank | N-gram | Count |
|---|---|---|
| 1 | Χ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ |
115,165 |
| 2 | Χ©Χ ΧΧΧΧ Χ 20 |
24,487 |
| 3 | Χ©ΧΧΧ ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ ΧΧΧ Χ |
19,413 |
| 4 | ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ ΧΧΧ Χ ΧΧͺΧΧΧΧ |
19,413 |
| 5 | ΧΧͺ ΧΧΧ€Χ’Χͺ ΧΧΧΧΧ¨Χ Χ©ΧΧ |
16,388 |
5-grams (Word):
| Rank | N-gram | Count |
|---|---|---|
| 1 | Χ©ΧΧΧ ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ ΧΧΧ Χ ΧΧͺΧΧΧΧ |
19,413 |
| 2 | Χ’Χ¨Χ ΧΧͺ ΧΧΧ€Χ’Χͺ ΧΧΧΧΧ¨Χ Χ©ΧΧ |
11,486 |
| 3 | ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ Χ©ΧΧΧ ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ |
10,724 |
| 4 | Χ©ΧΧΧΧΧ Χ©ΧΧΧ ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ ΧΧΧ Χ |
10,724 |
| 5 | ΧΧΧͺ ΧΧ ΧΧΧ¨ΧΧ Χ©Χ ΧΧ¨Χ¦ΧΧͺ ΧΧΧ¨ΧΧͺ |
7,604 |
2-grams (Subword):
| Rank | N-gram | Count |
|---|---|---|
| 1 | _ Χ |
39,073,833 |
| 2 | Χͺ _ |
29,026,407 |
| 3 | _ Χ |
24,932,558 |
| 4 | Χ _ |
24,128,474 |
| 5 | Χ _ |
21,592,884 |
3-grams (Subword):
| Rank | N-gram | Count |
|---|---|---|
| 1 | Χ Χ _ |
13,358,320 |
| 2 | Χ Χͺ _ |
11,186,966 |
| 3 | Χͺ _ Χ |
8,271,610 |
| 4 | _ Χ© Χ |
6,687,390 |
| 5 | Χ© Χ _ |
5,737,360 |
4-grams (Subword):
| Rank | N-gram | Count |
|---|---|---|
| 1 | _ Χ© Χ _ |
5,452,714 |
| 2 | _ Χ Χͺ _ |
2,964,460 |
| 3 | Χ Χͺ _ Χ |
2,726,223 |
| 4 | _ Χ’ Χ _ |
2,650,017 |
| 5 | Χ Χ Χ _ |
2,272,182 |
5-grams (Subword):
| Rank | N-gram | Count |
|---|---|---|
| 1 | _ Χ© Χ _ Χ |
1,545,782 |
| 2 | _ Χ Χ Χ _ |
1,326,505 |
| 3 | _ Χ Χͺ _ Χ |
1,316,470 |
| 4 | Χ _ Χ© Χ _ |
1,085,085 |
| 5 | Χ _ Χ© Χ _ |
843,378 |
Key Findings
- Best Perplexity: 2-gram (subword) with 388
- Entropy Trend: Decreases with larger n-grams (more predictable)
- Coverage: Top-1000 patterns cover ~13% of corpus
- Recommendation: 4-gram or 5-gram for best predictive performance
3. Markov Chain Evaluation
Results
| Context | Variant | Avg Entropy | Perplexity | Branching Factor | Unique Contexts | Predictability |
|---|---|---|---|---|---|---|
| 1 | Word | 1.1002 | 2.144 | 22.34 | 2,985,722 | 0.0% |
| 1 | Subword | 0.8730 | 1.831 | 7.49 | 25,039 | 12.7% |
| 2 | Word | 0.3737 | 1.296 | 2.25 | 66,677,134 | 62.6% |
| 2 | Subword | 0.6573 | 1.577 | 4.43 | 187,480 | 34.3% |
| 3 | Word | 0.1205 | 1.087 | 1.25 | 150,136,299 | 87.9% |
| 3 | Subword | 0.6833 | 1.606 | 3.99 | 829,497 | 31.7% |
| 4 | Word | 0.0427 π | 1.030 | 1.07 | 187,719,110 | 95.7% |
| 4 | Subword | 0.6743 | 1.596 | 3.51 | 3,312,743 | 32.6% |
Generated Text Samples (Word-based)
Below are text samples generated from each word-based Markov chain model:
Context Size 1:
Χ©Χ ΧΧΧ¦ΧΧ’ ΧΧΧ ΧΧ€Χ‘ΧΧ§ΧΧ ΧΧ’ΧΧΧ ΧΧΧΧΧ ΧΧ Χ¨ΧΧ Χ©Χ’ΧΧ¨Χ ΧΧ€ΧΧ¨ΧΧΧ Χ ΧΧΧΧ ΧΧ©Χ¨Χ Χ¨ΧΧΧΧͺ ΧΧΧ§ΧΧΧ€Χ‘ Χ©Χ ΧΧ©ΧΧ¨ ΧΧ€Χ©ΧΧΧΧͺ Χ‘ΧΧ ΧΧΧΧ£ ΧΧΧ ΧΧΧΧ ΧΧΧͺ ΧΧΧΧ¨ ΧΧ€Χ‘Χ§Χ ΧΧΧ ΧΧ¨ΧΧ ΧΧΧ‘ΧΧ£ ΧΧΧΧΧ’ Χ©ΧΧ ΧΧΧ ΧΧ’Χ Χ©ΧΧΧ ΧΧ¨ΧΧΧΧΧΧΧΧͺ ΧΧΧΧ ΧΧ’Χ ΧΧ¦ΧΧ ΧΧ©ΧΧΧΧ ΧΧΧΧΧ§ ΧΧΧΧ¨ΧΧΧͺ Χ¦Χ¨Χ€ΧͺΧΧͺ Χ’Χ¨ΧΧΧͺ ΧΧ€ΧΧ Χ ΧΧͺΧ ΧΧΧ¨ΧΧ ΧΧΧΧ¨ΧΧ ΧΧΧΧ§ΧΧ¨ ΧΧΧΧΧ Χ ΧΧΧ§Χ ΧΧΧ¦ΧΧ ΧͺΧ©ΧΧΧΧͺ
Context Size 2:
Χ’Χ ΧΧΧ ΧΧΧ©Χ ΧΧΧΧ ΧΧΧ‘ΧΧ¨ ΧΧͺ Χ©ΧΧΧΧ ΧΧΧ¨ΧΧ¨ ΧΧ¨ΧΧΧ ΧΧΧ¨Χ¨Χ Χ‘ ΧΧΧ Χ Χ©Χ ΧΧ’ΧΧ¨ Χ ΧΧ€ΧΧ ΧΧΧ Χ¨ΧΧΧΧ ΧΧΧΧͺΧ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ ΧΧΧΧΧ ΧΧΧ§Χ¨ΧΧΧ Χ ΧΧ’ΧΧͺ Χ§ΧΧ Χ‘ΧΧ€Χ¨Χ ΧΧΧ ΧΧ ΧΧ¨ ΧΧ§ΧΧΧsing unto godΧΧΧ ΧΧ ΧΧ¨ Χ‘ΧΧ€Χ¨...ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ ΧΧΧΧ¨ΧΧ Χ‘Χ’ΧΧΧΧΧͺ ΧΧΧ’ΧΧΧ Χ ΧΧΧΧ¨ΧΧ ΧΧΧΧΧ¨ ΧΧΧ¨ΧΧΧ‘ΧΧ Χ©ΧΧ’ΧΧ¨ΧΧ§ Χ’Χ Χ§ΧΧΧΧΧͺ ΧΧΧ Χ©ΧΧΧ ΧΧ£ ΧΧΧͺΧ¨ ΧΧΧ‘ΧΧ¨ΧΧ‘...
Context Size 3:
Χ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ Χ§Χ ΧΧΧ ΧΧΧΧ¨Χ ΧΧΧΧΧΧͺΧΧͺ ΧΧ¨ΧΧ ΧΧΧΧ¨Χ ΧΧΧΧΧΧͺΧΧͺ ΧΧΧΧΧΧ ΧΧΧΧ¨Χ ΧΧΧΧΧΧͺΧΧͺ ΧΧΧΧΧΧ ...ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ Χ§ΧΧΧ ΧΧ’ ΧΧΧΧΧΧΧΧΧ Χ¦ ΧΧΧΧΧ ΧΧΧͺ ΧͺΧ§Χ©ΧΧ¨Χͺ Χ¦ ΧΧΧΧΧ ΧΧ ΧΧΧΧΧΧΧΧ Χ¦ ΧΧΧΧΧ ΧΧ Χ§ΧΧΧ ΧΧ’ ΧΧΧΧΧΧΧ...Χ©Χ ΧΧ¨Χ¦ΧΧͺ ΧΧΧ¨ΧΧͺ ΧΧΧͺΧΧ‘Χ‘ Χ’Χ Χ‘Χ§Χ¨ΧΧ Χ’Χ ΧΧ§Χ¨Χ§Χ’ ΧΧ’Χ ΧͺΧ¦ΧΧΧΧ ΧΧΧΧΧ¨ Χ©Χ¦ΧΧΧΧ ΧΧΧΧΧ‘Χ ΧΧ©ΧΧΧͺ ΧΧΧ§Χ¨ ΧΧΧ ΧΧΧ¨Χ§ΧΧΧͺ ΧΧΧ¨ΧΧΧΧͺ...
Context Size 4:
Χ§ΧΧ©ΧΧ¨ΧΧ ΧΧΧ¦ΧΧ ΧΧΧ ΧΧ’Χ¨ΧΧͺ Χ©ΧΧΧΧΧ ΧΧ‘ΧΧ¨ Χ’ΧΧΧͺΧ ΧΧΧΧΧ ΧΧ ΧΧΧΧ ΧΧ ΧΧΧΧ ΧΧΧΧ¦Χ ΧΧΧ©Χ Χ©Χ ΧΧΧΧΧ©Χ ΧΧΧΧ Χ 20 ΧΧ Χ€ΧΧ§Χ ΧΧ ΧΧΧͺ ΧΧ Χ¨Χ©ΧΧ ΧΧΧ‘ΧΧ¨ ΧΧΧΧ¨Χ‘Χ Χ’Χ©Χ¨ΧΧͺ ΧΧΧ¨ΧΧͺ ΧΧΧ©Χ¨ΧΧ ΧΧΧ ΧΧ©ΧΧ¨ ΧΧΧΧΧΧΧ ΧΧΧ Χ€Χ§Χ ΧΧ¨ΧΧ©ΧΧ Χ ΧΧΧ...Χ©ΧΧΧ ΧͺΧΧ ΧΧͺ ΧΧ¨ΧΧΧ ΧΧ§Χ ΧΧΧ Χ ΧΧͺΧΧΧΧ Χ€ΧΧΧΧ§ΧΧΧΧ ΧΧ‘Χ¨Χ ΧΧΧΧΧ Χ©Χ ΧΧ’ΧΧΧΧ ΧΧ©ΧΧΧΧΧ
Generated Text Samples (Subword-based)
Below are text samples generated from each subword-based Markov chain model:
Context Size 1:
_Χ’ΧΧΧΧΧ_Χ§Χ¨ΧΧ¨ΧΧΧ_ΧΧ§Χ¨Χ._Χ§ΧΧΧΧ_ΧΧΧΧΧΧΧΧͺΧΧ¨ΧΧΧΧΧ’Χ_Χ©ΧΧ
Context Size 2:
_ΧΧ¨ΧΧΧ₯,_ΧΧΧ._ΧΧΧΧͺΧͺ_ΧΧ_ΧΧ©ΧΧΧ_ΧΧΧΧΧ§__ΧΧΧ¨Χ,_ΧΧͺ_ΧΧ/Χ Χ§Χ¨Χͺ
Context Size 3:
ΧΧ_ΧΧΧΧ Χ_ΧͺΧΧΧ_Χ¨Χ§ΧΧ‘ΧΧͺ_ΧΧ¦ΧΧ_ΧΧ©Χ¨ΧΧΧΧ€ΧΧ¨ΧΧͺ_ΧΧΧ§ΧΧΧ Χͺ_45_ΧΧΧΧΧ
Context Size 4:
_Χ©Χ_ΧΧΧΧ)_Χ©ΧΧΧΧ₯_ΧΧΧ_ΧΧͺ_ΧΧΧ_ΧΧΧΧΧΧ_ΧΧΧ¨ΧΧΧͺ_ΧΧ¨ΧΧ©ΧΧ_ΧΧΧ©ΧΧΧ_Χ‘Χ
Key Findings
- Best Predictability: Context-4 (word) with 95.7% predictability
- Branching Factor: Decreases with context size (more deterministic)
- Memory Trade-off: Larger contexts require more storage (3,312,743 contexts)
- Recommendation: Context-3 or Context-4 for text generation
4. Vocabulary Analysis
Statistics
| Metric | Value |
|---|---|
| Vocabulary Size | 1,343,537 |
| Total Tokens | 218,728,300 |
| Mean Frequency | 162.80 |
| Median Frequency | 5 |
| Frequency Std Dev | 6864.53 |
Most Common Words
| Rank | Word | Frequency |
|---|---|---|
| 1 | Χ©Χ | 5,459,894 |
| 2 | ΧΧͺ | 2,971,688 |
| 3 | Χ’Χ | 2,703,880 |
| 4 | ΧΧΧ | 1,339,510 |
| 5 | Χ’Χ | 1,154,254 |
| 6 | Χ | 905,656 |
| 7 | ΧΧ©Χ Χͺ | 775,632 |
| 8 | Χ | 760,765 |
| 9 | ΧΧ | 682,600 |
| 10 | ΧΧΧ | 665,182 |
Least Common Words (from vocabulary)
| Rank | Word | Frequency |
|---|---|---|
| 1 | markomannen | 2 |
| 2 | traditiones | 2 |
| 3 | possessionesque | 2 |
| 4 | bisterem | 2 |
| 5 | ΧΧ ΧΧΧΧΧΧΧ | 2 |
| 6 | Χ§Χ¨ΧΧΧΧΧͺΧΧ ΧΧ | 2 |
| 7 | Χ§Χ¨ΧΧΧΧΧͺΧΧΧΧΧ | 2 |
| 8 | ΧΧ ΧΧΧ¨ΧΧ₯ | 2 |
| 9 | Χ‘Χ§Χ‘ΧΧ€ΧΧΧ | 2 |
| 10 | ΧΧ‘Χ§Χ‘ΧΧ€ΧΧΧ | 2 |
Zipf's Law Analysis
| Metric | Value |
|---|---|
| Zipf Coefficient | 0.8691 |
| RΒ² (Goodness of Fit) | 0.995091 |
| Adherence Quality | excellent |
Coverage Analysis
| Top N Words | Coverage |
|---|---|
| Top 100 | 18.7% |
| Top 1,000 | 39.8% |
| Top 5,000 | 60.2% |
| Top 10,000 | 69.8% |
Key Findings
- Zipf Compliance: RΒ²=0.9951 indicates excellent adherence to Zipf's law
- High Frequency Dominance: Top 100 words cover 18.7% of corpus
- Long Tail: 1,333,537 words needed for remaining 30.2% coverage
5. Word Embeddings Evaluation
5.1 Cross-Lingual Alignment
5.2 Model Comparison
| Model | Dimension | Isotropy | Semantic Density | Alignment R@1 | Alignment R@10 |
|---|---|---|---|---|---|
| mono_32d | 32 | 0.8057 | 0.3812 | N/A | N/A |
| mono_64d | 64 | 0.7873 | 0.2918 | N/A | N/A |
| mono_128d | 128 | 0.7406 | 0.2357 | N/A | N/A |
| aligned_32d | 32 | 0.8057 π | 0.3678 | 0.1680 | 0.6000 |
| aligned_64d | 64 | 0.7873 | 0.2944 | 0.3600 | 0.7620 |
| aligned_128d | 128 | 0.7406 | 0.2283 | 0.4900 | 0.8080 |
Key Findings
- Best Isotropy: aligned_32d with 0.8057 (more uniform distribution)
- Semantic Density: Average pairwise similarity of 0.2999. Lower values indicate better semantic separation.
- Alignment Quality: Aligned models achieve up to 49.0% R@1 in cross-lingual retrieval.
- Recommendation: 128d aligned for best cross-lingual performance
6. Morphological Analysis (Experimental)
This section presents an automated morphological analysis derived from the statistical divergence between word-level and subword-level models. By analyzing where subword predictability spikes and where word-level coverage fails, we can infer linguistic structures without supervised data.
6.1 Productivity & Complexity
| Metric | Value | Interpretation | Recommendation |
|---|---|---|---|
| Productivity Index | 5.000 | High morphological productivity | Reliable analysis |
| Idiomaticity Gap | -0.772 | Low formulaic content | - |
6.2 Affix Inventory (Productive Units)
These are the most productive prefixes and suffixes identified by sampling the vocabulary for global substitutability patterns. A unit is considered an affix if stripping it leaves a valid stem that appears in other contexts.
Productive Prefixes
| Prefix | Examples |
|---|---|
-Χ |
ΧΧ€Χ¨ΧΧ‘Χ, ΧΧΧΧΧͺ, ΧΧΧ‘ΧΧΧΧ ΧΧΧ |
-Χ |
ΧΧΧΧΧΧ, ΧΧΧ¨ΧΧ¨ΧΧΧ¦ΧΧ, ΧΧΧΧΧΧΧΧ¨ΧΧ |
-Χ |
ΧΧΧΧΧ, ΧΧΧΧΧ₯, ΧΧ¨Χ’Χ©Χ |
-Χ |
ΧΧΧ Χ¦, ΧΧΧ¨ΧΧΧ ΧΧ§ΧΧͺ, ΧΧΧΧΧ¦Χͺ |
-Χ |
ΧΧ‘Χ€Χ§Χ, ΧΧΧ‘ΧΧΧ¨Χ, ΧΧΧΧ¨ΧΧ€ΧΧ |
-Χ© |
Χ©ΧΧΧΧΧ¨Χ£, Χ©ΧΧͺΧΧΧ, Χ©ΧΧΧΧ¨ |
-ΧΧ |
ΧΧΧ‘ΧΧΧΧ ΧΧΧ, ΧΧΧ¨ΧΧΧΧ, ΧΧΧ¨ΧΧ‘Χͺ |
-Χ |
ΧΧΧΧΧΧΧΧΧ, ΧΧ ΧΧΧ€ΧΧ‘ΧΧ€ΧΧΧΧ€ΧΧΧΧͺ, ΧΦΆΧ¦Φ°ΧΦ°ΦΌΧ’ΧΦΉΧ Φ΄Χ |
Productive Suffixes
| Suffix | Examples |
|---|---|
-Χ |
ΧΧΧ‘ΧΧΧΧ ΧΧΧ, ΧΧΧΧΧΧΧΧ¨ΧΧ, ΧΧΧΧ¨ΧΧ€ΧΧ |
-Χ |
ΧΧΧΧΧ, ΧΧ€Χ¨ΧΧ‘Χ, ΧΧΧ¨ΧΧ¨ΧΧΧ¦ΧΧ |
-Χͺ |
Χ ΧΧΧΧΧͺ, ΧΧΧΧΧͺ, ΧΧ ΧΧΧ€ΧΧ‘ΧΧ€ΧΧΧΧ€ΧΧΧΧͺ |
-ΧΧ |
ΧΧΧ‘ΧΧΧΧ ΧΧΧ, ΧΧΧΧΧΧΧΧ¨ΧΧ, ΧΧΧΧ¨ΧΧ€ΧΧ |
-ΧΧͺ |
Χ ΧΧΧΧΧͺ, Χ€Χ¨Χ§ΧΧΧΧΧΧΧͺ, ΧΧΧ¨ΧΧΧ ΧΧ§ΧΧͺ |
-Χ |
ΧΧΧΧΧΧΧΧΧ, ΧΧΧΧ Χ‘Χ§Χ, ΧΧ‘Χ€Χ§Χ |
-Χ |
ΧΧ¨ΧΧΧΧΧ©ΧΧ, ΧΧΧΧΧΧ, ΧΧΧΧΧ |
-s |
lugares, wootens, hijras |
6.3 Bound Stems (Lexical Roots)
Bound stems are high-frequency subword units that are semantically cohesive but rarely appear as standalone words. These often correspond to the 'core' of a word that requires inflection or derivation to be valid.
| Stem | Cohesion | Substitutability | Examples |
|---|---|---|---|
ΧͺΧ€Χ§Χ |
2.54x | 314 contexts | ΧͺΧ€Χ§ΧΧ’, ΧΧͺΧ€Χ§Χ, ΧͺΧ€Χ§ΧΧ¨ |
ΧΧ€ΧΧ’ |
2.45x | 92 contexts | ΧΧ€ΧΧ’Χ, ΧΧΧ€ΧΧ’, ΧΧΧ€ΧΧ’ |
ΧΧΧΧ |
2.81x | 51 contexts | ΧΧΧΧΧ, ΧΧΧΧΧ, ΧΧΧΧΧ |
Χ’ΧΧΧ |
1.93x | 275 contexts | Χ’ΧΧΧΧͺ, Χ’ΧΧΧΧ, ΧΧ’ΧΧΧ |
ΧΧ¨ΧΧ |
2.21x | 126 contexts | ΧΧ¨ΧΧ Χ, ΧΧ¨ΧΧ Χ, ΧΧ¨ΧΧ Χ |
ΧΧ¦ΧΧ |
2.23x | 120 contexts | ΧΧΧ¦ΧΧ Χ, ΧΧΧ¦ΧΧ Χ, Χ§ΧΧ¦ΧΧ Χ |
ΧͺΧ§ΧΧ€ |
2.13x | 149 contexts | ΧͺΧ§ΧΧ€Χͺ, ΧΧͺΧ§ΧΧ€, ΧͺΧ§ΧΧ€Χ |
ΧΧΧΧ |
1.90x | 259 contexts | ΧΧΧΧ Χ, ΧΧΧΧ Χͺ, ΧΧΧΧ Χ¦ |
Χ§ΧΧΧ |
1.95x | 203 contexts | Χ§ΧΧΧΧ, Χ§ΧΧΧΧ, Χ§ΧΧΧΧͺ |
ΧΧΧ¨Χ€ |
1.73x | 292 contexts | ΧΧΧ¨Χ€Χ, ΧΧΧ¨Χ€Χ, ΧΧΧ¨Χ€Χ |
ΧͺΧΧΧ |
1.69x | 272 contexts | ΧͺΧΧΧ Χ, ΧͺΧΧΧ Χ, ΧͺΧΧΧ Χ |
Χ¨Χ‘ΧΧ |
2.40x | 45 contexts | ΧΧ¨Χ‘ΧΧ, Χ¨Χ‘ΧΧΧ, ΧΧ¨Χ‘ΧΧ |
6.4 Affix Compatibility (Co-occurrence)
This table shows which prefixes and suffixes most frequently co-occur on the same stems, revealing the 'stacking' rules of the language's morphology.
| Prefix | Suffix | Frequency | Examples |
|---|---|---|---|
-Χ |
-Χͺ |
158 words | ΧΧΧ€ΧΧ‘ΧΧͺ, ΧΧΧΧΧΧΧΧΧͺ |
-Χ |
-Χͺ |
158 words | ΧΧΧΧ¨ΧΧΧΧ Χͺ, ΧΧΧΧ©ΧΧΧΧͺ |
-Χ |
-Χ |
154 words | ΧΧΧΧΧ¨ΧΧ, ΧΧΧΧΧΧ¨ΧΧ |
-Χ |
-Χ |
144 words | ΧΧ ΧΧΧΧ‘Χ, ΧΧΧ¨Χ¦ΧΧ€ΧΧ |
-Χ |
-ΧΧ |
136 words | ΧΧΧΧΧ¨ΧΧ, ΧΧΧΧΧΧ¨ΧΧ |
-Χ |
-Χ |
114 words | ΧΧͺΧ¨ΧΧ§ΧΧ, ΧΧΧ Χ¨Χ |
-Χ |
-ΧΧ |
110 words | ΧΧΧ¨Χ¦ΧΧ€ΧΧ, ΧΧΧΧΧ‘ΧΧΧ |
-Χ |
-ΧΧͺ |
105 words | ΧΧΧΧ©ΧΧΧΧͺ, ΧΧ¨Χ¦ΧΧΧ ΧΧΧΧͺ |
-Χ |
-Χ |
90 words | ΧΧΧΧ ΧΧΧ, ΧΧΧ€ΧΧ¨Χ§ΧΧ |
-Χ |
-Χͺ |
85 words | ΧΧΧ§ΧΧΧ¨ΧΧΧͺ, ΧΧͺΧ’Χ©Χ¨Χͺ |
6.5 Recursive Morpheme Segmentation
Using Recursive Hierarchical Substitutability, we decompose complex words into their constituent morphemes. This approach handles nested affixes (e.g., prefix-prefix-root-suffix).
| Word | Suggested Split | Confidence | Stem |
|---|---|---|---|
| sipstrassi | sipstras-s-i |
7.5 | s |
| ΧΧΧ¨ΧΧΧ Χ©ΧΧΧ¨ | ΧΧΧ¨ΧΧΧ Χ©Χ-Χ-Χ¨ |
7.5 | Χ |
| ΧΧΧ ΧΧΧ ΧΧΧΧ Χ©ΧΧΧ¨ | ΧΧΧ ΧΧΧ ΧΧΧΧ Χ©Χ-Χ-Χ¨ |
7.5 | Χ |
| ΧΧΧ‘ΧΧ ΧΧ¨ΧΧΧ | ΧΧ-Χ‘ΧΧ ΧΧ¨Χ-ΧΧ |
6.0 | Χ‘ΧΧ ΧΧ¨Χ |
| ΧΧͺΧΧ ΧΧ§ΧΧͺΧΧΧ | ΧΧͺΧΧ ΧΧ§ΧΧͺ-ΧΧ-Χ |
6.0 | ΧΧͺΧΧ ΧΧ§ΧΧͺ |
| Χ©ΧΧΧ€Χ©Χ¨ΧΧͺΧ | Χ©Χ-ΧΧ€Χ©Χ¨ΧΧͺ-Χ |
6.0 | ΧΧ€Χ©Χ¨ΧΧͺ |
| ΧΧ©ΧͺΧ§Χ€ΧΧΧΧͺΧΧΧ | ΧΧ©ΧͺΧ§Χ€ΧΧΧΧͺ-ΧΧ-Χ |
6.0 | ΧΧ©ΧͺΧ§Χ€ΧΧΧΧͺ |
| ΧΧ€Χ¨ΧΧΧͺΧΧΧ | ΧΧ€Χ¨ΧΧΧͺ-ΧΧ-Χ |
6.0 | ΧΧ€Χ¨ΧΧΧͺ |
| ΧΧΧΧ¨ΧΧΧΧΧΧΧΧ | ΧΧ-ΧΧ¨ΧΧΧΧΧΧ-ΧΧ |
6.0 | ΧΧ¨ΧΧΧΧΧΧ |
| ΧΧͺΧΧΧΧ©ΧΧͺΧ | ΧΧͺΧΧΧΧ©-ΧΧͺ-Χ |
6.0 | ΧΧͺΧΧΧΧ© |
| Χ’Χ§Χ¨ΧΧ ΧΧͺΧΧΧ | Χ’Χ§Χ¨ΧΧ ΧΧͺ-ΧΧ-Χ |
6.0 | Χ’Χ§Χ¨ΧΧ ΧΧͺ |
| Χ©ΧΧΧΧΧ¨ΧΧΧͺ | Χ©Χ-ΧΧΧΧ¨Χ-ΧΧͺ |
6.0 | ΧΧΧΧ¨Χ |
| ΧΧ¨ΧΧ©ΧΧ ΧΧΧͺΧ | ΧΧ¨ΧΧ©ΧΧ Χ-ΧΧͺ-Χ |
6.0 | ΧΧ¨ΧΧ©ΧΧ Χ |
| ΧΧΧΧΧΧͺΧΧΧ | ΧΧΧΧΧΧͺ-ΧΧ-Χ |
6.0 | ΧΧΧΧΧΧͺ |
| ΧΧ€Χ ΧΧΧΧΧΧ | Χ-Χ€Χ ΧΧΧΧ-ΧΧ |
6.0 | Χ€Χ ΧΧΧΧ |
6.6 Linguistic Interpretation
Automated Insight: The language Hebrew shows high morphological productivity. The subword models are significantly more efficient than word models, suggesting a rich system of affixation or compounding.
7. Summary & Recommendations
Production Recommendations
| Component | Recommended | Rationale |
|---|---|---|
| Tokenizer | 64k BPE | Best compression (4.19x) |
| N-gram | 2-gram | Lowest perplexity (388) |
| Markov | Context-4 | Highest predictability (95.7%) |
| Embeddings | 100d | Balanced semantic capture and isotropy |
Appendix: Metrics Glossary & Interpretation Guide
This section provides definitions, intuitions, and guidance for interpreting the metrics used throughout this report.
Tokenizer Metrics
Compression Ratio
Definition: The ratio of characters to tokens (chars/token). Measures how efficiently the tokenizer represents text.
Intuition: Higher compression means fewer tokens needed to represent the same text, reducing sequence lengths for downstream models. A 3x compression means ~3 characters per token on average.
What to seek: Higher is generally better for efficiency, but extremely high compression may indicate overly aggressive merging that loses morphological information.
Average Token Length (Fertility)
Definition: Mean number of characters per token produced by the tokenizer.
Intuition: Reflects the granularity of tokenization. Longer tokens capture more context but may struggle with rare words; shorter tokens are more flexible but increase sequence length.
What to seek: Balance between 2-5 characters for most languages. Arabic/morphologically-rich languages may benefit from slightly longer tokens.
Unknown Token Rate (OOV Rate)
Definition: Percentage of tokens that map to the unknown/UNK token, indicating words the tokenizer cannot represent.
Intuition: Lower OOV means better vocabulary coverage. High OOV indicates the tokenizer encounters many unseen character sequences.
What to seek: Below 1% is excellent; below 5% is acceptable. BPE tokenizers typically achieve very low OOV due to subword fallback.
N-gram Model Metrics
Perplexity
Definition: Measures how "surprised" the model is by test data. Mathematically: 2^(cross-entropy). Lower values indicate better prediction.
Intuition: If perplexity is 100, the model is as uncertain as if choosing uniformly among 100 options at each step. A perplexity of 10 means effectively choosing among 10 equally likely options.
What to seek: Lower is better. Perplexity decreases with larger n-grams (more context). Values vary widely by language and corpus size.
Entropy
Definition: Average information content (in bits) needed to encode the next token given the context. Related to perplexity: perplexity = 2^entropy.
Intuition: High entropy means high uncertainty/randomness; low entropy means predictable patterns. Natural language typically has entropy between 1-4 bits per character.
What to seek: Lower entropy indicates more predictable text patterns. Entropy should decrease as n-gram size increases.
Coverage (Top-K)
Definition: Percentage of corpus occurrences explained by the top K most frequent n-grams.
Intuition: High coverage with few patterns indicates repetitive/formulaic text; low coverage suggests diverse vocabulary usage.
What to seek: Depends on use case. For language modeling, moderate coverage (40-60% with top-1000) is typical for natural text.
Markov Chain Metrics
Average Entropy
Definition: Mean entropy across all contexts, measuring average uncertainty in next-word prediction.
Intuition: Lower entropy means the model is more confident about what comes next. Context-1 has high entropy (many possible next words); Context-4 has low entropy (few likely continuations).
What to seek: Decreasing entropy with larger context sizes. Very low entropy (<0.1) indicates highly deterministic transitions.
Branching Factor
Definition: Average number of unique next tokens observed for each context.
Intuition: High branching = many possible continuations (flexible but uncertain); low branching = few options (predictable but potentially repetitive).
What to seek: Branching factor should decrease with context size. Values near 1.0 indicate nearly deterministic chains.
Predictability
Definition: Derived metric: (1 - normalized_entropy) Γ 100%. Indicates how deterministic the model's predictions are.
Intuition: 100% predictability means the next word is always certain; 0% means completely random. Real text falls between these extremes.
What to seek: Higher predictability for text generation quality, but too high (>98%) may produce repetitive output.
Vocabulary & Zipf's Law Metrics
Zipf's Coefficient
Definition: The slope of the log-log plot of word frequency vs. rank. Zipf's law predicts this should be approximately -1.
Intuition: A coefficient near -1 indicates the corpus follows natural language patterns where a few words are very common and most words are rare.
What to seek: Values between -0.8 and -1.2 indicate healthy natural language distribution. Deviations may suggest domain-specific or artificial text.
RΒ² (Coefficient of Determination)
Definition: Measures how well the linear fit explains the frequency-rank relationship. Ranges from 0 to 1.
Intuition: RΒ² near 1.0 means the data closely follows Zipf's law; lower values indicate deviation from expected word frequency patterns.
What to seek: RΒ² > 0.95 is excellent; > 0.99 indicates near-perfect Zipf adherence typical of large natural corpora.
Vocabulary Coverage
Definition: Cumulative percentage of corpus tokens accounted for by the top N words.
Intuition: Shows how concentrated word usage is. If top-100 words cover 50% of text, the corpus relies heavily on common words.
What to seek: Top-100 covering 30-50% is typical. Higher coverage indicates more repetitive text; lower suggests richer vocabulary.
Word Embedding Metrics
Isotropy
Definition: Measures how uniformly distributed vectors are in the embedding space. Computed as the ratio of minimum to maximum singular values.
Intuition: High isotropy (near 1.0) means vectors spread evenly in all directions; low isotropy means vectors cluster in certain directions, reducing expressiveness.
What to seek: Higher isotropy generally indicates better-quality embeddings. Values > 0.1 are reasonable; > 0.3 is good. Lower-dimensional embeddings tend to have higher isotropy.
Average Norm
Definition: Mean magnitude (L2 norm) of word vectors in the embedding space.
Intuition: Indicates the typical "length" of vectors. Consistent norms suggest stable training; high variance may indicate some words are undertrained.
What to seek: Relatively consistent norms across models. The absolute value matters less than consistency (low std deviation).
Cosine Similarity
Definition: Measures angular similarity between vectors, ranging from -1 (opposite) to 1 (identical direction).
Intuition: Words with similar meanings should have high cosine similarity. This is the standard metric for semantic relatedness in embeddings.
What to seek: Semantically related words should score > 0.5; unrelated words should be near 0. Synonyms often score > 0.7.
t-SNE Visualization
Definition: t-Distributed Stochastic Neighbor Embedding - a dimensionality reduction technique that preserves local structure for visualization.
Intuition: Clusters in t-SNE plots indicate groups of semantically related words. Spread indicates vocabulary diversity; tight clusters suggest semantic coherence.
What to seek: Meaningful clusters (e.g., numbers together, verbs together). Avoid over-interpreting distances - t-SNE preserves local, not global, structure.
General Interpretation Guidelines
- Compare within model families: Metrics are most meaningful when comparing models of the same type (e.g., 8k vs 64k tokenizer).
- Consider trade-offs: Better performance on one metric often comes at the cost of another (e.g., compression vs. OOV rate).
- Context matters: Optimal values depend on downstream tasks. Text generation may prioritize different metrics than classification.
- Corpus influence: All metrics are influenced by corpus characteristics. Wikipedia text differs from social media or literature.
- Language-specific patterns: Morphologically rich languages (like Arabic) may show different optimal ranges than analytic languages.
Visualizations Index
| Visualization | Description |
|---|---|
| Tokenizer Compression | Compression ratios by vocabulary size |
| Tokenizer Fertility | Average token length by vocabulary |
| Tokenizer OOV | Unknown token rates |
| Tokenizer Total Tokens | Total tokens by vocabulary |
| N-gram Perplexity | Perplexity by n-gram size |
| N-gram Entropy | Entropy by n-gram size |
| N-gram Coverage | Top pattern coverage |
| N-gram Unique | Unique n-gram counts |
| Markov Entropy | Entropy by context size |
| Markov Branching | Branching factor by context |
| Markov Contexts | Unique context counts |
| Zipf's Law | Frequency-rank distribution with fit |
| Vocab Frequency | Word frequency distribution |
| Top 20 Words | Most frequent words |
| Vocab Coverage | Cumulative coverage curve |
| Embedding Isotropy | Vector space uniformity |
| Embedding Norms | Vector magnitude distribution |
| Embedding Similarity | Word similarity heatmap |
| Nearest Neighbors | Similar words for key terms |
| t-SNE Words | 2D word embedding visualization |
| t-SNE Sentences | 2D sentence embedding visualization |
| Position Encoding | Encoding method comparison |
| Model Sizes | Storage requirements |
| Performance Dashboard | Comprehensive performance overview |
About This Project
Data Source
Models trained on wikipedia-monthly - a monthly snapshot of Wikipedia articles across 300+ languages.
Project
A project by Wikilangs - Open-source NLP models for every Wikipedia language.
Maintainer
Citation
If you use these models in your research, please cite:
@misc{wikilangs2025,
author = {Kamali, Omar},
title = {Wikilangs: Open NLP Models for Wikipedia Languages},
year = {2025},
doi = {10.5281/zenodo.18073153},
publisher = {Zenodo},
url = {https://huggingface.co/wikilangs}
institution = {Omneity Labs}
}
License
MIT License - Free for academic and commercial use.
Links
- π Website: wikilangs.org
- π€ Models: huggingface.co/wikilangs
- π Data: wikipedia-monthly
- π€ Author: Omar Kamali
- π€ Sponsor: Featherless AI
Generated by Wikilangs Models Pipeline
Report Date: 2026-01-13 14:18:23



















